<?xml version="1.0" encoding="UTF-8"?><oembed><type>video</type><version>1.0</version><html>&lt;iframe src=&quot;https://www.loom.com/embed/23e2f87a2bc94fb598d324bb12ce626b&quot; frameborder=&quot;0&quot; width=&quot;1238&quot; height=&quot;928&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>928</height><width>1238</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>928</thumbnail_height><thumbnail_width>1238</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/23e2f87a2bc94fb598d324bb12ce626b-498ea7188ff9fdbd.gif</thumbnail_url><duration>119.957</duration><title>How Pact Ensures Policy and Trust</title><description>This Loom explains how Pact works as a policy, authority, contract, and trust framework for an AI echo worker within an ambiguous team workspace. The creator describes Pact using a constitution of general rules that cannot be bypassed and a contract of specific, task-level rules. They then demonstrate activating a mission to search for a voice AI pilot using CompilotKit Atlassian Agent with tasks like compare, cleaning, code, documentation, and solutions. Because the task is ambiguous, requests can be escalated to a human in the loop for approval and execution as needed.</description></oembed>